Regulatory Open Forum

 View Only
  • 1.  Bioburden Sample Size

    This message was posted by a user wishing to remain anonymous
    Posted 29-Jul-2020 09:11
    This message was posted by a user wishing to remain anonymous


    Is there a recommended number of samples to test for bio burden under ISO 11737-1?

    We are doing a bio burden test under ISO 11737-1. The lab suggests to ask Regulatory about the number of samples to be tested per batch. I would have thought the sample would have to be scientifically justified by Engineering with QA. The item is a balloon for an endorectal procedure.

    Thank you

  • 2.  RE: Bioburden Sample Size

    Posted 30-Jul-2020 02:52
    This is something that the company I work for will have to start doing soon as we have just about exited the V&V phase for 3 of our devices being developed. I am interested to hear what the more experienced regulatory professionals would say.

    My thoughts, for now, are as follows:
    • Establish an environmental monitoring/bioburden procedure
    • Create a test plan including all your devices
    • Understand the production throughput
    • Consider a statistical sampling method to select devices for Bioburden release testing

    The procedure will establish critical and alert limits but depending on your device and sterilisation technique these can be varied.

    I have probably only figured out a small percentage of the actual requirements but this is where I am at. I hope it helps a little and stimulates further discussion.


    Malcolm Applewhite
    QA/RA Manager
    Cape Town
    South Africa

  • 3.  RE: Bioburden Sample Size

    Posted 30-Jul-2020 04:48

    Malcolm you have raised some really valid points which definitely should be considered and/or incorporated into the manufacturing/testing processes within the product realisation activities.  To answer the question about the sample size, there are quite a factors to be taken into consideration for bioburden activities.  Some of the factors to consider are:

    • Risk classification of the device including type of device
    • Type of device, exposure to bioburdens, hard to reach places, bacteriostasis/fungistasis (USP <71>)
    • Intended purpose of device such as surface contacting versus implant and sterilisation processes
    • Duration of contact, which is directly related to above and sterilisation processes
    • Sterilisation process - there are some information contained in sterilisation standards regarding sampling
    • Lot or batch size, number of batches (per month/per year)
    • Lot release requirements, some require bioburden testing for batch release

    If you read through 11737 and some of the other related sterilisation standards will give some ideas on the number of samples to take.  There is also other information related such as the MPN calculation, bioburden calculation, and FDA has some information in the Analytical Manual about microbiological testing with some sample size tables (this is for food and cosmetics, but gives some ideas for devices).  In simple terms, I have seen sample sizes from n = 3 to n = 10 per lot.  You could have some fun doing some analysis, calculations, and statistics, but the most important is whatever sample size you determine make sure you clearly document reasons and justifications for the sample size.

    Richard Vincins RAC
    Vice President Global Regulatory Affairs

  • 4.  RE: Bioburden Sample Size

    Posted 10-Aug-2020 11:08
    I agree with the previous comments – also note that, per Richard's statement that he usually sees sample sizes of 3–10, this is called out in ISO 11737-1:2018, A.8.1 as "common practice" for routine monitoring. 

    In addition to the considerations Richard has listed, as you build a bioburden data set, your own historical data can help justify the sample size you use (i.e., if your bioburden has been consistently low, you may be able to more easily justify using N=3 as opposed to N=10).

    Jennifer Cabralda, RAC (US, CAN)
    Director, Regulatory Affairs
    Richmond, BC, Canada

  • 5.  RE: Bioburden Sample Size

    Posted 11-Aug-2020 09:11
    Dear Jennifer,

    Thanks for the reference to the standard. I have definitely read through this Annex before but have no recollection of the sample size. I don't think that I will forget that now.

    I would like to add that although I agree that historical data can justify your sample size, I do not agree that low bioburden values result in low sample sizes. I would argue that minimal deviation in bioburden values (low or high) would result in a smaller sample size. i.e. if your bioburden data is consistent you could justify a smaller sample size.


    Malcolm Applewhite
    QA/RA Manager
    Cape Town
    South Africa

  • 6.  RE: Bioburden Sample Size

    Posted 11-Aug-2020 18:01

    Sample size determination in this case is an interested process. The problem starts with the concept of sterilization. In my simple way of thinking, sterilization "kills bugs". Start with a known concentration, the bioburden, and apply the sterilization method to achieve a Sterility Assurance Level, SAL, of six. Each level is a power of ten reduction. The concept is that the sterilization parameters (depending on the sterilization method) will "kill the required proportion of bugs". If so, there are not enough "bugs" left to harm to patient.

    Often, the sterilization method includes the initial sampling plan. For example, ISO 11737-1:2018 A.8.1 says that the ISO 11137 series provides the information for sterilization by radiation.

    The assumption in ISO 11737-1:2018 is that the initial bioburden level is known based on the standards that apply to the sterilization method. ISO 11737-1:2018 asks whether the bioburden changed.

    The requirement is a monitoring plan with two major elements: sampling plan and sampling frequency.

    Put simply, frequency is "how often do I check". This depends on the stability of the process that creates the bioburden. More stable processes allow less frequent checking. A good rule of thumb establishes the initial frequency to match the particle count frequency from the ISO 14644 family. Use dynamic adjustment methods based on the results.

    Put simply, sampling is "how do I check". The sampling plan is trying to detect a change from the initial bioburden to the current bioburden. This is a hypothesis test. ISO 11737-1:2018 A.8.1 identifies three factors to help determine the sample size:
    Tradition (3 to 10)
    The change in bioburden to detect
    The variation in the number of viable microorganisms present on individual items (between items)
    Also, the standard uses the concept of SIP in which the variability is on an item (within items)

    This is a standard, small sample, hypothesis test (using Student's t-distribution). Mind common issues in a hypothesis test such as the alpha and beta risks to determine the sample size.

    Dan O'Leary CQA, CQE
    Swanzey NH
    United States